Cost alerts¶
Cost alerts tell your team when spend crosses a line, so a runaway bill reaches you before the invoice does. An alert watches a cost figure against a threshold and, when it trips, sends a message to your notification channels.
Creating an alert¶
An alert has a few simple parts:
| Setting | What it means |
|---|---|
| Threshold amount | The cost level that triggers the alert |
| Currency | CHF or EUR, so the threshold matches how you think about the bill |
| Scope | A single project, or the whole organisation |
| Enabled | A toggle to activate or pause the alert without deleting it |
Create as many alerts as you need, for example a per-project ceiling for each team plus one organisation-wide safety net.
How alerts are evaluated¶
Alerts are checked two ways:
- On a schedule: a background job re-evaluates every alert at a fixed interval (every few minutes by default), so you do not need to be looking at the app.
- On manual refresh: when you refresh cost data from the dashboard, alerts for that organisation are re-evaluated immediately.
When the watched cost is at or above the threshold and the alert is enabled, it fires.
What happens when an alert fires¶
- A message is sent to every enabled notification channel (Slack, Teams, Discord, Pushover, email).
- The event is written to the alert history, with the date, the cost value that tripped it, the message and the delivery status.
- The alert's last triggered date is updated, which you can see in the alert list.
Throttling
Alerts are throttled so a threshold that stays crossed does not spam your channels repeatedly. You get told when the situation starts, not on every check.
Budgets and the 80% warning¶
Beyond hard threshold alerts, OpenWatch tracks budget burn. On the dashboard, a budget that has consumed more than 80% of its allotment is highlighted as a warning, giving you a heads-up before the limit is actually reached.
Good practices¶
- Set thresholds slightly below the pain point, so the alert is a warning and not a post-mortem.
- Use per-project alerts to give each team ownership of its own spend, and one org-wide alert as a backstop.
- Make sure at least one notification channel is configured and tested, otherwise an alert can fire with nowhere to go.
Where to go next¶
- Configure where alerts are delivered: Notification channels.
- Understand the numbers alerts watch: Cost dashboard.
- Reduce the spend that trips alerts: rightsizing and orphan resources.